健康体检女性含超声和血清标志物的乳腺恶性结节Nomogram预测模型的构建与验证  

Construction and validation of a nomogram prediction model for malignant breast nodules based on ultrasound and serum markers in women undergoing health check-ups

作　　者：刘琰 梁雯雯 王鹏昌 LIU Yan;LIANG Wen-wen;WANG Peng-chang(Health Management Center,Jiaozuo Second People's Hospital(the First Affiliated Hospital of Henan Polytechnic University),Jiaozuo 454000,Henan,CHINA)

机构地区：[1]焦作市第二人民医院(河南理工大学第一附属医院)健康管理中心,河南焦作454000

出　　处：《海南医学》2025年第5期623-629,共7页Hainan Medical Journal

基　　金：河南省医学科技攻关计划项目(编号:LHGJ20200829)。

摘　　要：目的基于超声和血清标志物构建健康体检女性乳腺恶性结节Nomogram预测模型,并对所构建的模型进行验证。方法选取2021年1月至2024年5月河南理工大学第一附属医院(焦作市第二人民医院)收治的491例乳腺结节患者作为研究对象,按照7∶3比例分为建模集344例和验证集147例,根据乳腺恶性结节患病情况将建模集患者分为恶性组102例和良性组242例。比较两组患者的临床相关资料,采用单因素、LASSO-Logistic回归分析筛选出乳腺结节患者发生恶性病变的独立相关影响因素,构建Nomogram预测模型,采用受试者工作特征(ROC)曲线、校准曲线评估该模型的预测效能,并将验证集资料纳入模型进行验证。结果恶性组患者的乳腺癌家族史患者占比、乳腺超声图像的影像报告与数据系统(BI-RADS-US)分类>3类患者占比、糖类抗原153(CA153)阳性患者占比、铁蛋白阳性患者占比、癌胚抗原(CEA)阳性患者占比分别为7.84%、88.24%、60.78%、89.22%、67.65%,明显高于良性组的1.24%、26.03%、6.20%、20.25%、10.74%,差异均有统计学意义(P<0.05);恶性组患者的结节直径、癌胚抗原相关细胞黏附分子1(CEACAM1)分别为(2.94±0.97)cm、(568.29±172.66)ng/mL,明显长(高)于良性组的(1.98±0.65)cm、(370.59±121.64)ng/mL,差异均有统计学意义(P<0.05);LASSO-Logistic多因素回归分析结果显示,BI-RADS-US分类>3类、CEACAM1升高、CA153阳性、铁蛋白阳性、CEA阳性均是乳腺恶性结节的独立相关影响因素(P<0.05);基于上述相关影响因素构建的乳腺恶性结节预测模型的预测风险一致性指数(C-index)为0.977;ROC曲线、校准曲线显示,该模型准确性及预测效能较高。结论乳腺恶性结节的独立相关影响因素为BI-RADS-US分类>3类、CEACAM1升高、CA153阳性、铁蛋白阳性、CEA阳性。基于此构建的乳腺恶性结节预测模型准确性、预测效能较高,可精准识别早期的潜在高风险人群,为健康�Objective To construct a nomogram prediction model for malignant breast nodules based on ultrasound and serum markers in women undergoing health check-ups,and to validate the constructed model.Methods A total of 491 patients with breast nodules admitted to the First Affiliated Hospital of Henan Polytechnic University(Jiaozuo Second People's Hospital)from January 2021 to May 2024 were selected as the study subjects.The patients were divided into a modeling set(344 cases)and a validation set(147 cases)at a ratio of 7:3.The patients in the modeling set were further categorized into a malignant group(102 cases)and a benign group(242 cases)based on the prevalence of malignant breast nodules.The clinical data of the two groups were compared.Univariate and LASSO-Logistic regression analysis were employed to identify independent factors associated with the development of malignant lesions in breast nodules.A nomogram prediction model was constructed,and the predictive performance of the model was evaluated using receiver operating characteristic(ROC)curves and calibration curves.The validation set data were incorporated into the model for verification.Results The proportions of patients with a family history of breast cancer,patients with BI-RADS-US classification>3,patients with positive carbohydrate antigen 153(CA153),patients with positive ferritin,and patients with positive carcinoembryonic antigen(CEA)in the malignant group were 7.84%,88.24%,60.78%,89.22%,and 67.65%,respectively,which were significantly higher than 1.24%,26.03%,6.20%,20.25%,and 10.74%in the benign group(P<0.05).The nodule diameter and carcinoembryonic antigen-related cell adhesion molecule 1(CEACAM1)levels in the malignant group were(2.94±0.97)cm and(568.29±172.66)ng/mL,respectively,which were significantly longer(higher)than(1.98±0.65)cm and(370.59±121.64)ng/mL in the benign group(P<0.05).LASSO-Logistic multivariate regression analysis showed that BI-RADS-US classification>3,elevated CEACAM1,positive CA153,positive ferritin,and positive CEA wer

关 键 词：健康体检女性 乳腺恶性结节 NOMOGRAM 模型构建 模型验证 影响因素 

分 类 号：R711[医药卫生—妇产科学]